Amino acid racemization in Quaternary foraminifera from the … · 2020-02-28 · 2008; O’Regan...

Geochronology 1 53ndash67 2019httpsdoiorg105194gchron-1-53-2019copy Author(s) 2019 This work is distributed underthe Creative Commons Attribution 40 License

Amino acid racemization in Quaternary foraminifera from theYermak Plateau Arctic OceanGabriel West1 Darrell S Kaufman2 Francesco Muschitiello3 Matthias Forwick4 Jens Matthiessen5Jutta Wollenburg5 and Matt OrsquoRegan1

1Department of Geological Sciences Stockholm University 10691 Stockholm Sweden2School of Earth and Sustainability Northern Arizona University Flagstaff AZ 86011 USA3Department of Geography University of Cambridge Cambridge CB2 3EN UK4Department of Geosciences UiT The Arctic University of Norway 9037 Tromsoslash Norway5Alfred Wegener Institute for Polar and Marine Research 27570 Bremerhaven Germany

Correspondence Gabriel West (gabrielwestgeosuse)

Received 11 June 2019 ndash Discussion started 16 July 2019Accepted 25 October 2019 ndash Published 18 November 2019

Abstract Amino acid racemization (AAR) geochronol-ogy is a powerful tool for dating Quaternary marine sedi-ments across the globe yet its application to Arctic Oceansediments has been limited Anomalous rates of AAR inforaminifera from the central Arctic were reported in pre-viously published studies indicating that either the rate ofracemization is higher in this area or inaccurate age modelswere used to constrain the sediment ages This study inves-tigates racemization rates in foraminifera from three well-dated sediment cores taken from the Yermak Plateau duringthe 2015 TRANSSIZ (TRansitions in the Arctic SeasonalSea Ice Zone) expedition on RV Polarstern D and L iso-mers of the amino acids aspartic acid (Asp) and glutamic acid(Glu) were separated in samples of the planktic foraminiferNeogloboquadrina pachyderma and the benthic species Cas-sidulina neoteretis to quantify the extent of racemization Intotal 241 subsamples were analysed extending back to ma-rine oxygen isotope stage (MIS) 7 Two previously publishedpower functions which relate the extent of racemization ofAsp and Glu in foraminifera to sample age are revisited anda comparison is made between the ages predicted by thesecalibrated age equations and independent geochronologicalconstraints available for the cores Our analyses reveal an ex-cellent match between ages predicted by a global compilationof racemization rates for N pachyderma and confirm that aproposed Arctic-specific calibration curve is not applicable atthe Yermak Plateau These results generally support the ratesof AAR determined for other cold bottom water sites and fur-

ther highlight the anomalous nature of the purportedly highrate of racemization indicated by previous analyses of centralArctic sediments

1 Introduction

Dating Quaternary marine sediments from the Arctic Oceanhas been a long-standing problem and a number of stud-ies (eg Backman et al 2004 Stein 2011 Alexandersonet al 2014) highlight the challenges of establishing firmchronologies for these sediments Assigning ages for the var-ious lithostratigraphic units in Arctic Ocean sediments ishowever of paramount importance as the development ofaccurate age models is key to contextualize Arctic palaeo-ceanography within Earthrsquos climate system

Amino acid racemization (AAR) geochronology was firstapplied specifically to Arctic Ocean sediments in the pioneer-ing studies of Sejrup et al (1984) and later in that of Mackoand Aksu (1986) Sejrup et al (1984) used AAR results tochallenge the prevailing view of slow (mm kyrminus1) sedimen-tation rates (e g Clark 1970) in the Arctic and warned thatthe existing age interpretations derived from bio- and mag-netostratigraphy could be substantially flawed On the otherhand Macko and Aksu (1986) found that AAR chronologyof sediments from the Alpha Ridge supported the acceptedages established by the use of these two dating techniquesthus supporting arguments for slow sedimentation rates inthis region Due to the limitations ndash associated with small

Published by Copernicus Publications on behalf of the European Geosciences Union

54 G West et al Amino acid racemization in Quaternary foraminifera from the Yermak Plateau

sample size and low temporal resolution ndash of these early stud-ies the application of AAR chronology for Arctic Ocean sed-iments decreased in the following years However major de-velopments in AAR dating such as the increased use of themore rapidly racemizing aspartic acid over isoleucine (egGoodfriend et al 1996) and the application of the reversephase liquid chromatography to determine amino acid DLvalues in small samples (Kaufman and Manley 1998) cou-pled with improvements in Arctic Ocean sediment chronos-tratigraphies (eg Jakobsson et al 2001 Backman et al2008 OrsquoRegan et al 2008) have reignited the interest inAAR studies during the past decade

Kaufman et al (2008) related the extent of AAR in thepolar foraminifera species Neogloboquadrina pachydermacollected from various parts of the Arctic Ocean (MendeleevRidge Lomonosov Ridge Northwind Ridge) to sedimentages by developing a calibrated age equation for the past150 kyr This was followed by the study of Kaufman etal (2013) who reported the results of AAR analysis of mul-tiple foraminifera species from sediment cores taken in theArctic Atlantic and Pacific oceans and defined a generalrate of racemization at deep-sea sites Both models relied oncalibration of the rate of racemization of aspartic (Asp) andglutamic (Glu) acids by independent dating methods (such asradiocarbon dating and correlations with the orbitally tunedstacked oxygen isotope record Martinson et al 1987) yetfor the past 150 kyr these two age equations produced sig-nificantly different ages for a given DL value The globalcompilation of AAR showed that the rate of racemization inthe Arctic Ocean was higher than expected in samples of Npachyderma older than 35 ka and this could not be explainedby differences in the rate of AAR in N pachyderma rela-tive to other species It remains to be seen whether the ob-served rates of racemization are anomalously high comparedto other ocean basins or alternatively the age constraints forthe studied Arctic Ocean cores are inaccurate Sedimentarysequences with robust age control and satisfactory preserva-tion of carbonate microfossils from the Arctic are required tofurther investigate this question

Here we present the results of amino acid racemiza-tion analysis of planktic (N pachyderma) and benthic (Cneoteretis) foraminifera from samples of three well-datedsediment cores from the Yermak Plateau The rates of racem-ization in the two species are compared and the Arctic-specific (Kaufman et al 2008) and global (Kaufman et al2013) calibrated age equations are used to estimate the agesof the samples The calculated ages are compared with thosederived from age models constructed by combining 14C dat-ing oxygen isotope stratigraphy and the correlation of envi-ronmental magnetic properties to the global benthic oxygenisotope record

Figure 1 Locations of the studied cores and water depth at coresites Basemap IBCAO Jakobsson et al (2012)

2 Materials and methods

21 Sample material and study area

The analysed foraminifera samples were taken from threesediment cores (PS920039-2 PS920045-2 and PS9254-1)recovered from the Yermak Plateau during the TRANSSIZ(TRansitions in the Arctic Seasonal Sea Ice Zone) expedi-tion (19 May to 28 June 2015) on the research icebreaker RVPolarstern (PS92) (Table 1 and Fig 1)

The Yermak Plateau located north of the Svalbardarchipelago lies at the gateway to the central Arctic Oceanin an area characterized by the interaction between Atlanticand Arctic waters The cold surface layer is dominated by thepolar foraminifer N pachyderma which has highest abun-dances in the top 200 m (Carstens and Wefer 1992 Carstenset al 1997 Pados and Spielhagen 2014 Greco et al 2019)Cassidulina neoteretis is one of the dominant benthic speciesin sediments from the Yermak Plateau (Bergsten 1994 Wol-lenburg and Mackensen 1998 Wollenburg et al 2004)and its abundance is often associated with the presence ofAtlantic water (Polyak and Solheim 1994 Slubowska etal 2005)

The cores were recovered from water depths between 915and 1464 m ndash in the Upper Polar Deep Water ndash far be-low the core of the warmer Atlantic layer which typicallysits between depths of 200 and 500 m (Jones 2001) Po-tential temperature and salinity profiles from nearby CTD(conductivityndashtemperaturendashdepth) stations illustrate the re-

Geochronology 1 53ndash67 2019 wwwgeochronologynet1532019

G West et al Amino acid racemization in Quaternary foraminifera from the Yermak Plateau 55

Table 1 Sediment cores analysed in this study

Latitude Longitude Water Core Bottom waterCore Core type (N) (E) depth (m) length (m) temperature (C)

PS9239-2 kastenlot core 8194983 1382833 1464 8600 minus076PS9245-2 gravity core 8189333 9768833 915 5195 minus025PS9254-1 gravity core 81123 8466167 1145 4145 minus043

Figure 2 Potential temperature and salinity profiles of CTD sta-tions (PS9239-8 46-2 and 55-1) nearest to the coring sites CTDstations 46-2 and 55-1 did not sample the water depths of cores 45-2and 54-1 thus the markers approximate where they would sit on theCTD profiles

lationship between the coring sites and the local watermasses (Fig 2)

22 Sample ages and quantification of age modeluncertainties

The age model of core PS9239-2 was initially developed byKremer et al (2018) who used a combination of AMS 14Cdates sedimentological correlations and the occurrence ofthe benthic foraminifer Pullenia bulloides as a stratigraphicmarker for event 51 (sim 81 ka) in the polar North Atlantic(Haake and Pflaumann 1989) and Arctic Ocean (Wollenburget al 2001) This age model was further developed by Wierset al (2019) by recalibrating the 14C dates integrating ad-ditional carbon and oxygen isotope data from foraminiferaand applying a correlation between environmental magneticparameters (ratio of anhysteretic remanent susceptibility andbulk magnetic susceptibility) with the global stacked benthicforaminifera δ18OLR record of Lisiecki and Raymo (2005)

Wiers et al (2019) extended this method to cores PS9245-2and PS9254-1 (Figs 3 and 4) The ability to correlate en-vironmental magnetic parameters to the global benthic δ18Orecord in late Quaternary sediments from the Yermak Plateauwas first proposed by Xuan et al (2012) who also utilized14C dating and oxygen isotope measurements on plankticforaminifera

To more rigorously provide a measure of the uncertaintyinherent to the correlation-based age models we applied anautomated Monte Carlo algorithm for proxy-to-proxy strati-graphical alignment based on the assumption that changesin magnetic parameters (kARMk ratio of anhysteretic re-manent susceptibility to bulk magnetic susceptibility) in ourmarine cores and variations in the δ18OLR stack are virtuallysynchronous The algorithm used here consists of a modi-fied and improved routine that builds upon previous work byMalinverno (2013) and Muschitiello et al (2015) and Mus-chitiello (2016) and that has been widely used for synchro-nization of different palaeoclimate archives (eg Wohlfarthet al 2018 Muschitiello et al 2019) The method was de-signed to obtain a sample of optimal alignment functions thatrelates depth in one record to age (or depth) in another Themedian of the samples gives the best correlation and theirvariability provides a measure of the uncertainty associatedwith the alignment whereby the alignment function will havelarger uncertainties where the match between the proxy timeseries is poorly constrained Further details on adjustment ofthe agendashdepth model of Wiers et al (2019) are contained inAppendix A

23 Sample preparation and analytical procedure

Individual specimens of the foraminifera species N pachy-derma and C neoteretis were picked from the gt 63 microm frac-tion of wet-sieved oven-dried (4 h at 30 C) 2 cm thick slicesof sediment samples from core depths in which tests of thetwo species were abundant The sediment slices were takenfrom u-channel samples that had been used for palaeomag-netic measurements implying that the samples came fromthe central parts of the cores This resulted in a total of 1010 and 5 sampled stratigraphic levels in cores PS9239-2PS9245-2 and PS9254-1 respectively Seven core depthswere sampled for both N pachyderma and C neoteretis incores PS9245-2 and PS9254-1 Core PS9239-2 was onlysampled for N pachyderma

wwwgeochronologynet1532019 Geochronology 1 53ndash67 2019


Figure 3 Lithostratigraphic correlation (on original depth scale) of sediment cores PS9239-2 PS9245-2 and PS9254-1 based on Wiers etal (2019) Dashed lines represent tie points pB bulk density

Figure 4 The agendashdepth models (Wiers et al 2019) of the coresin this study are based on correlation of environmental magneticparameters (kARMk) and the global benthic δ18O stack of Lisieckiand Raymo (2005)

The foraminifera tests were first cleaned by sonication (ap-proximately 1 s) in a bath sonicator then were immersed in1 mL of 3 H2O2 for 2 h Some samples required up to 4 hof immersion for a complete removal of organic matter Thetests were rinsed 3 times with reagent grade water (grade I)

and dried under a laminar flow hood In order to create sub-samples for each sampled core depth 7ndash12 tests were picked(minimum 7 tests for C neoteretis and minimum 10 for Npachyderma) and placed in micro-reaction hydrolysis vialspre-sterilized by heating at 550 C The number of subsam-ples within a sample was constrained by the number of avail-able tests and their level of preservation in a particular sam-ple To dissolve the tests 8 microL of 6 M HCl was added to thevials and the solution was sealed under N2 gas The subsam-ples were hydrolysed at 110 C for 6 h in order to convert theoriginal protein present in the tests into free amino acids Fol-lowing hydrolysis the subsamples were evaporated in a vac-uum desiccator A total of 4 microL of 001 M HCl with 10 microM L-homoarginine spike was added to the evaporate and the sub-samples were then injected onto a high-performance liquidchromatograph (HPLC) All measurements were performedat the Amino Acid Geochronology Laboratory Northern Ari-zona University The HPLC instrumentation and procedureused are described by Kaufman and Manley (1998) Inter-laboratory standards (Wehmiller 1984 Table 2) were anal-ysed to monitor instrument performance

The extent of amino acid racemization was determined byanalysing the peak-area ratio of D and L enantiomers (nineamino acids measured in total) Data analysis focused on Aspand Glu as these amino acids are abundant in foraminiferaprotein and among the best-resolved chromatographically(eg Kaufman et al 2013) and they provide the basis ofprevious calibrated age equations

3 Results

A total of 32 foraminifera samples (10 from core PS9239-213 from core PS9245-1 and 9 from core PS9254-1) sub-



Table 2 Summary of aspartic acid and glutamic acid mean DL values (following data screening) and independent ages (based on Wierset al 2019) of samples analysed in this study n number of subsamples analysed ex number of excluded subsamples Values in bold aremean DL values that show depth reversals Interlaboratory standards (ILC A B C) analysed during the course of this study are also shown

Depth Age Number DL DLLab ID Core (m bsf) (ka) n ex of tests Asp σ Glu σ

Neogloboquadrina pachyderma

15544 PS9239-2 020 67 7 6 10 0121 ndash 0056 ndash15545 PS9239-2 035 89 4 4 10 ndash ndash ndash ndash15546 PS9239-2 150 308 7 3 10 0234 0018 0082 001015547 PS9239-2 300 642 8 0 10 0258 0025 0103 002115548 PS9239-2 380 804 10 2 10 0233 0016 0095 001515549 PS9239-2 581 1263 9 5 10 0272 0063 0110 003515550 PS9239-2 651 1392 9 1 10 0313 0014 0153 001515551 PS9239-2 721 1508 10 6 10 0298 0016 0118 001615552 PS9239-2 831 1685 10 5 10 0283 0021 0111 001515553 PS9239-2 841 1701 10 4 10 0329 0017 0169 0013

16814 PS9245-2 120 298 9 3 10 0190 0012 0085 0009lowast16815 PS9245-2 159 395 11 0 10 0195 0008 0084 0007lowast16817 PS9245-2 312 901 8 4 10 0269 0019 0124 002716818 PS9245-2 413 1478 7 3 10 0345 0014 0176 0017lowast17531 PS9245-2 445 1718 11 0 12 0351 0027 0186 002916819 PS9245-2 473 1949 6 2 10 0362 0023 0189 0016

lowast16822 PS9254-1 179 401 7 1 10 0166 0011 0070 0009lowast16826 PS9254-1 466 1323 15 7 12 0395 0010 0224 0010lowast16827 PS9254-1 494 1358 6 1 10 0358 0035 0175 0040lowast16828 PS9254-1 51 1378 4 1 10 0345 0018 0186 0020

Cassidulina neoteretis

17315 PS9245-2 104 254 7 1 7 0121 0003 0035 000217316 PS9245-2 139 345 2 1 7 0159 ndash 0055 ndashlowast17317 PS9245-2 159 395 7 1 7 0167 0010 0055 000617318 PS9245-2 175 435 3 2 7 0189 ndash 0070 ndashlowast17319 PS9245-2 312 901 8 2 10 0274 0021 0103 000717320 PS9245-2 417 1508 7 3 7 0341 0017 0160 0017lowast17321 PS9245-2 445 1718 10 3 10 0277 0022 0154 0025

17322 PS9254-1 147 342 5 0 7 0175 0012 0068 0011lowast17323 PS9254-1 179 401 6 0 7 0154 0004 0055 0003lowast17324 PS9254-1 466 1323 7 2 7 0319 0012 0165 0023lowast17325 PS9254-1 494 1358 6 2 7 0326 0017 0168 0018lowast17326 PS9254-1 51 1378 5 0 7 0304 0012 0155 0015

Interlaboratory standards

ILC-A NA NA NA NA NA NA 0412 0225ILC-B NA NA NA NA NA NA 0718 0453ILC-C NA NA NA NA NA NA 0842 0845

lowast Samples marked with an asterisk are from core depths sampled for both Neogloboquadrina pachyderma and Cassidulina neoteretis NA notapplicable

divided into 241 subsamples (average 75 subsamples persample) were analysed The average number of foraminiferatests per subsample was 102 for N pachyderma and 75 forC neoteretis (Table 2) The results of amino acid analysis ofall subsamples are contained in Table S1 in the Supplement

In order to assess whether subsample size had a significantinfluence on the resulting DL values a further 47 subsam-ples of the foraminifer N pachyderma were taken each with20 individual tests from stratigraphic depths already sam-pled (subsamples with 10 tests) The results of this analysiswere then compared with 42 subsamples that comprised 10



tests No statistically significant difference could be observedat the 001 probability level (the details of this analysis arecontained in the Supplement)

Outliers were removed by following the screening processof Kosnik and Kaufman (2008)

1 Subsamples with L SerL Asp values greater than 08were excluded as this could be an indicator of contam-ination by modern amino acids

2 The positive covariance of Asp and Glu acid DL valuesin foraminifera is well known (eg Hearty et al 2004Kosnik and Kaufman 2008) subsamples that deviatedfrom the expected trend were excluded (Fig 5)

3 Remaining subsamples with DL Asp or DL Glu val-ues outside the 2σ of the sample mean were rejected toemphasize the central tendency of the data

This screening process resulted in the rejection of 68 sub-samples equivalent to 282 of all subsamples The overallrejection rate is higher than for most other studies as sum-marized by Kaufman et al (2013) although in this studysubsamples from core 39-2 account for over half (53 ) ofall rejected subsamples despite only 35 of all subsamplesbeing taken from this core Only 188 ndash typical of previousstudies ndash of the subsamples from core 45-2 was rejected andthis core provided 40 of all subsamples

Asp and Glu DL values show an increasing trend withdepth for both foraminifera species (Fig 6) in all coresDL values of Asp and Glu in core PS9239-2 are lower atequivalent depths below seafloor than compared to PS9245-2 and PS9254-1 when samples from below 3 m are consid-ered Eight of the 32 samples exhibit mean Asp DL valuesin reverse stratigraphic order (highlighted with bold in Ta-ble 2) Six of the reversed values overlap within 2σ errorswith the sample from shallower depths and only two meanAsp DL values from C neoteretis samples in core PS9245-2 (at 445 m) and in core PS9254-1 (at 179 m) lack such anoverlap Mean DL Glu values also exhibit stratigraphicallyreversed values in these samples with overlaps within theerror range of the overlying sample

Mean DL values of Asp and Glu are generally higherfor N pachyderma than for C neoteretis for samples fromthe same depth (Table 2) This difference can be observed inboth cores PS9245-2 and PS9254-1 It is difficult to quan-tify the differences in racemization rates between the twospecies as little data are available for an extensive com-parison and some stratigraphic levels suffer from reversedDL values However as a basic approximation a plot ofN pachyderma versus C neoteretis mean DL values im-plies that Asp racemizes approximately 17 faster in Npachyderma than in C neoteretis while Glu racemizes about23 ndash26 faster in N pachyderma (Fig 7)

Figure 5 Covariance of aspartic acid and glutamic acid DL valuesin all subsamples Open symbols indicate rejected results with therejection criterion indicated by the symbol colour

4 Discussion

41 Relation between DL values and depth andinter-specific differences

The extent of racemization generally increases with depth inboth N pachyderma and C neoteretis samples and this con-forms to the expected diagenetic behaviour of amino acids inforaminifera Eight samples have stratigraphically reversedmean DL Asp and Glu values but six of these values over-lap within the 2σ uncertainty envelope with the overlyingsample

For a given sediment depth the extent of racemization inN pachyderma samples is lower in PS9239-2 below 3 mthan in the other two cores (Fig 6) This is consistent withthe higher sedimentation rates in PS9239-2 below 3 m com-pared to the other two cores based on the independent agecontrol of Wiers et al (2019)

Differences in the rates of AAR (16 for Asp and 23 ndash26 for Glu ndash Fig 7) between N pachyderma and Cneoteretis tests from the same depths are similar to those doc-umented in other AAR studies (eg King and Neville 1977Kaufman et al 2013) Little is known about the extent andcauses of the differences in racemization rates in these twospecies but differences of similar magnitude between DLvalues of different taxa were previously reported in the lit-



Figure 6 Extent of racemization (mean DL in aspartic acid andglutamic acid) in samples of N pachyderma and C neoteretis plot-ted against depth in sediment cores PS9239-2 PS9245-2 andPS9254-1 Error bars represent plusmn1σ deviation from the samplemean Least square regression lines (power fit) are shown for allcores Data listed in Table 2

erature For example Kaufman et al (2013) found that Aspracemized 12 ndash16 faster in Pulleniatina obliquiloculatathan in N pachyderma As N pachyderma and C neoteretisare amongst the most commonly occurring planktic and ben-thic foraminifera species in the Arctic the quantification ofthe relative rates of racemization between the two specieswill aid future AAR studies in the Arctic and could be aug-mented by future laboratory heating experiments

Stratigraphically reversed DL Asp values which donot overlap within the 2σ uncertainty only appear in Cneoteretis samples One such reversal is in core PS9245-2 insample UAL17321 from 445 m core depth (total core length520 m) In this sample the mean DL Asp value of 0277(σ = 0022) is essentially equal to the mean value of 0274(σ = 0021) from 312 m depth It is difficult to explain theorigin of the stratigraphically reversed mean values in the Cneoteretis samples Sediment mixing (eg bioturbation re-worked material entrained in sea ice) offers one possible ex-planation and may account for reversals on decimetre scalesbut it seems less likely that metre-scale downward mixing ndashwhich would be required to explain the reversal discussed in

Figure 7 Extent of racemization (DL) for aspartic and glutamicacids in N pachyderma versus C neoteretis samples from coresPS9245-2 and PS9254-1 with linear regression and line of equal-ity

PS9245-2 ndash occurs Icebergs are capable of extensive sed-iment reworking but shipboard parasound and multibeamdata show that the coring sites were free from iceberg scour-ing The excellent correlation between the records precludesthe possibility of large erosional events that could have re-moved or reworked sediments

However the small number (7) of foraminifera tests usedin subsamples of C neoteretis could also make the DL val-ues more susceptible to variations in individual tests Forexample Lougheed et al (2018) recently highlighted thelarge heterogeneity in the age distribution of foraminiferaobtained from discrete depth intervals using 14C dating ofsingle foraminifera

42 Relation between DL values and sample age

A biplot of DL values against sample ages reveals that asexpected the rate of Asp and Glu racemization is higher inyounger samples and decreases with sample age as the re-action progresses towards equilibrium (Fig 8) DL valuesincrease in a predictable manner over time as shown by leastsquare regression power curves fitted to the mean DL Aspand Glu values for each of the cores An exceptionally goodfit for N pachyderma samples is achieved in core PS9245-2 (with R2 values of 099 and 098 for DL Asp and Glurespectively for individual regression lines see Fig S1) InC neoteretis samples both Asp and Glu appear to racemizeat similar rates in cores PS9245-2 and PS9254-1 DissimilarDL Asp and Glu value between samples of comparable agesfrom different cores may originate from differences in sedi-mentation rates between the cores variable post-depositionalenvironmental factors including differences in geothermalgradients or variable diagenetic processes These are dis-cussed in more detail below

43 Palaeotemperature and other possible effects

In monospecific foraminifera samples the rate of racemiza-tion is often considered to be primarily controlled by the inte-



Figure 8 Mean DL values of aspartic and glutamic acids plottedagainst the independent ages of Wiers et al (2019) for N pachy-derma and C neoteretis samples from cores PS9239-2 PS9245-2and PS9254-1 Ages predicted (based on DL Asp and Glu values)by the globally calibrated age equation of Kaufman et al (2013)and the Arctic-specific age equation (Kaufman et al 2008) are alsodisplayed for reference Error bars represent plusmn1σ deviation fromthe mean DL value of the sample and age uncertainty

grated post-depositional temperature (Kaufman et al 2013)Therefore if the DL value and age of a sample are knownthe AAR age equation can be solved for palaeotemperaturesHere we provide a rudimentary estimation of the differencein the effective diagenetic temperatures required to accountfor the offset between known-age-equivalent DL values be-tween the coring sites To achieve this we relied on the tem-perature equation originally derived by Kaufman (2006) us-ing simple power law kinetics for Asp in the foraminiferaspecies Pulleniatina obliquiloculata and increased the DLvalues of our N pachyderma samples by 16 to account forthe lower rate of racemization in this taxon as establishedby Kaufman et al (2013) For comparison between the threecore sites we determined the approximate DL Asp valuespredicted by power-fit functions for each core at 150 ka thenthe effective diagenetic temperatures integrated over the past150 kyr were derived from the temperature equation The ab-solute derived temperatures need to be interpreted with cau-tion as no formal assessment of the racemization kineticsin N pachyderma and an assessment of the uncertainties as-

Table 3 Approximate effective diagenetic temperatures at the threecore sites integrated over the past 150 kyr (C)

Estimated effective Approximate averagediagenetic temperatures sedimentation rate

Core over the past 150 kyr (C) during MIS 5 (cm kyrminus1)

PS9239-2 +42plusmn 08 45PS9245-2 +58plusmn 08 2PS9254-1 +82plusmn 08 19

Figure 9 Comparison of sample ages for the two foraminifera taxaas predicted by the globally calibrated age equation of Kaufmanet al (2013) based on DL Asp values and ages assigned to thesamples based on the age model developed by Wiers et al (2019)with line of equality

sociated with these were completed Furthermore there area number of complicating factors to these temperature esti-mates that are discussed below However relative tempera-ture differences between the sites are likely still meaning-ful The results imply that temperature differences (1T ) ofsim 16ndash40 C are required to account for the observed differ-ences in racemization rates between the cores over the past150 kyr (Table 3) Details on palaeotemperature calculationsare contained in the Supplement

It is important to emphasize that AAR palaeotemperatureestimates can only provide a measure of the long-term av-erage post-depositional temperature rather than the temper-ature of the water during calcification The estimated differ-ence in effective diagenetic temperatures between PS9239-2and PS9254-1 (4plusmn16 C) cannot be easily accounted for bychanges in bottom water temperatures alone Modern bottomwater temperatures at the three coring locations are relativelysimilar (between minus025 and minus076 C) (Table 1) althoughwe do not know how similar these remained across glacialcycles Presently the coring sites are beyond the reach of thewarm (gt 05 C) Atlantic layer (Fig 2) which would needto be at significantly greater depths to generate such a largespread in the effective diagenetic temperatures of the coringsites It is unclear how this could have happened in the past



and we do not see it as a plausible explanation for the ap-parent differences in calculated effective diagenetic temper-atures

Apart from differences in bottom water temperatures dif-ferences in the geothermal gradient among sites could alsoimpact the palaeotemperature signal Unfortunately no directmeasurements of heat flow were obtained at the coring sta-tions during TRANSSIZ and existing heat flow data from theYermak Plateau are relatively sparse (Okay and Crane 1993Shephard et al 2018) Existing measurements do indicatethat regions of the Yermak Plateau are characterized by rela-tively high heat flow (gt 100 mW mminus2) but with considerablespatial variability Most measurements lie in the range of 50to 100 mW mminus2 (Shephard et al 2018) If we use a thermalconductivity of 117 W mKminus1 reported from measurementson a core obtained from the Yermak Plateau (Shephard et al2018) heat flow of 50ndash100 m Wmminus2 translates into geother-mal gradients of 43 to 85 C kmminus1 (0043 to 0085 C mminus1)At a depth of 6 m below the seafloor this range of geothermalgradients would amount to a present-day in situ temperaturedifference of only 025 C Thus it also seems unlikely thatlocal heat flow variations would significantly affect rates ofracemization at the shallow sediment depth of these samples

Sejrup and Haugen (1994) have theorized that variationsin post-depositional microbial activity could have a profoundimpact on AAR and that the extent of microbial activity is re-lated to sedimentation rates Shells deposited in settings withlow sedimentation rates could have been exposed longer tomicrobial activity at the seabed than shells that were buriedrapidly and therefore more rapidly removed from the tapho-nomically active zone Such increased exposure could re-sult in a higher degree of protein and amino acid degrada-tion in the early diagenetic phase giving rise to higher initialDL values The sedimentation rate in core PS9239-2 is esti-mated to be around 45 cm kyrminus1 during marine oxygen iso-tope stage (MIS) 5 (assigned to the depth interval between35ndash6 m) 25 times that of the sedimentation rate at the othertwo core sites (Table 3) The comparatively lower DL valuesin core PS9239-2 could reflect the faster burial and therebybetter preservation of indigenous proteins in tests from the35ndash6 m depth interval yet the exact mechanisms of micro-bially mediated racemization remain unresolved Currentlywe lack sufficient data to more fully explore the reasons forinter-site offsets in the AAR rates

44 Comparing AAR age models

In order to assess the validity of the previously derivedArctic-specific (Kaufman et al 2008) and globally cali-brated (Kaufman et al 2013) age equations at the YermakPlateau ages predicted by the two models were generatedusing the mean DL Asp and Glu values from N pachy-derma and C neoteretis tests in this study (Table 4) Noages using the Arctic-specific age equation were generatedfor C neoteretis samples as this calibration is based on AAR

in N pachyderma tests only Ages predicted by the modelswere plotted against the corresponding DL values and dis-played for reference in the biplot of DL values against sam-ple age (Fig 8) Age uncertainties were derived using theintra-sample variability (plusmn1σ ) propagated through the ageequations

Modelled ages generated using the globally calibrated ageequation of Kaufman et al (2013) and mean DL Asp val-ues in N pachyderma are most similar to the sample agesof Wiers et al (2019) (Fig 8) The Arctic-specific age equa-tion of Kaufman et al (2008) results in significantly youngerages than those based on the Wiers et al (2019) age model(Fig 8) This implies that the Arctic-specific age equationcannot be used to constrain the ages of sediments from theYermak Plateau

These results generally support the rates of AAR deter-mined for other deep-sea (cold bottom water) sites furtherhighlighting the peculiar high rate of racemization indicatedby previous analyses of central Arctic sediments For exam-ple previously published (Kaufman et al 2008) mean DLAsp values for N pachyderma samples from a central Arcticcore from the Lomonosov Ridge (9612-1PC) are around 04for samples dated to 85ndash123 ka The globally calibrated ageequation predicts approximate ages of 263 kyr for this DLAsp value (with plusmn158 kyr error assuming an uncertainty of6 typical of mean DL values in this study) and sampleswith similar DL values from the Yermak Plateau would dateto around 232plusmn14 ka The considerable differences could in-dicate different post-depositional temperatures variable sed-imentation rates diagenetic processes issues with samplehandling storage and processing or may point towards er-rors in the existing age models of central Arctic sedimentsFurther studies are clearly required to investigate these op-tions

AAR ages based on DL Asp values and the globally cal-ibrated AAR curve strongly correlate with the independentages of Wiers et al (2019) for both species (Fig 9) howeverthere is a significant scatter in AAR ages older thansim 130 kaExcluding all samples with stratigraphically reversed DLAsp values (regardless of whether there is an overlap within2σ errors or not) increases the correlation coefficient (r) to096 in N pachyderma and to 099 in C neoteretis samplesWhile well-correlated the globally calibrated age equationslightly (sim on average 2 ) underestimates the ages of Npachyderma samples with the highest level of agreement be-tween the ages observed in core PS9245-2 The ages of Cneoteretis samples are mostly underestimated ndash on averageby 32 This is expected considering the lower racemizationrates of C neoteretis observed in this study (approximately16 slower than N pachyderma) Both sets of data are cou-pled with high values of scatter with differences betweenthe modelled ages and the ages based on Wiers et al (2019)reaching over 50 of the sample age in some cases Never-theless the globally calibrated age equation provides a re-liable age approximation for our Arctic Ocean sediments



Table 4 Independent sample ages based on Wiers et al (2019) and sample ages and uncertainties based on the globally calibrated ageequation of Kaufman et al (2013) and the Arctic-specific age equation of Kaufman et al (2008) based on mean DL values of Asp and Glu

Age (ka) based Age Asp Age Asp Age Glu Age Gluon Wiers Kaufman Kaufman Kaufman Kaufman

Depth et al et al et al et al et alLab ID Core (m bsf) (2019) (2013) (ka) (2008) (ka) (2013) (ka) (2008) (ka)


15544 PS9239-2 020 67 67plusmn 73 142 15015546 PS9239-2 150 308 510plusmn 39 334plusmn 26 367plusmn 45 275plusmn 3415547 PS9239-2 300 642 688plusmn 67 419plusmn 41 647plusmn 132 397plusmn 8115548 PS9239-2 380 804 503plusmn 35 331plusmn 23 502plusmn 32 348plusmn 5515549 PS9239-2 581 1263 874plusmn 187 501plusmn 109 832plusmn 299 441plusmn 14015550 PS9239-2 651 1392 1281plusmn 56 667plusmn 29 1618plusmn 195 747plusmn 7315551 PS9239-2 721 1508 1126plusmn 57 606plusmn 31 1025plusmn 116 493plusmn 6715552 PS9239-2 831 1685 964plusmn 68 539plusmn 38 832plusmn 109 447plusmn 6015553 PS9239-2 841 1701 1463plusmn 75 737plusmn 38 1992plusmn 271 876plusmn 6716814 PS9245-2 120 298 269plusmn 17 207plusmn 13 401plusmn 42 292plusmn 3116815 PS9245-2 159 395 292plusmn 12 220plusmn 09 390plusmn 32 286plusmn 2416817 PS9245-2 312 901 903plusmn 55 514plusmn 33 1046plusmn 192 534plusmn 11616818 PS9245-2 413 1478 1676plusmn 68 817plusmn 33 2447plusmn 236 934plusmn 9017531 PS9245-2 445 1718 1862plusmn 136 883plusmn 65 3038plusmn 380 1021plusmn 15916819 PS9245-2 473 1949 1943plusmn 123 912plusmn 58 2921plusmn 247 1047plusmn 8916822 PS9254-1 179 401 181plusmn 12 154plusmn 10 257plusmn 36 214plusmn 2716826 PS9254-1 466 1323 2539plusmn 64 1115plusmn 28 4456plusmn 199 1374plusmn 6116827 PS9254-1 494 1358 2095plusmn 184 965plusmn 87 2960plusmn 436 926plusmn 21216828 PS9254-1 510 1378 1676plusmn 87 817plusmn 43 2808plusmn 302 1021plusmn 110


17315 PS9245-2 104 254 67plusmn 02 44plusmn 0317316 PS9245-2 139 345 156plusmn 136plusmn ndash17317 PS9245-2 159 395 181plusmn 11 136plusmn 1517318 PS9245-2 175 435 265plusmn 248plusmnndash17319 PS9245-2 312 901 827plusmn 63 647plusmn 4417320 PS9245-2 417 1508 1618plusmn 81 1931plusmn 20517321 PS9245-2 445 1718 855plusmn 68 1756plusmn 28517322 PS9254-1 147 342 209plusmn 14 230plusmn 3717323 PS9254-1 179 401 141plusmn 04 136plusmn 0717324 PS9254-1 466 1323 1319plusmn 50 2085plusmn 29117325 PS9254-1 494 1358 1409plusmn 73 2180plusmn 23417326 PS9254-1 510 1378 1138plusmn 45 1785plusmn 173

when DL Asp values from N pachyderma samples are uti-lized

5 Conclusions

The extent of racemization of Asp and Glu increases withincreasing age in both N pachyderma and C neoteretis sam-ples and is higher in N pachyderma than in C neoteretisfrom the same stratigraphic levels

The globally calibrated age equation of Kaufman etal (2013) provides a reliable age control for sediments fromthe Yermak Plateau The Arctic-specific calibrated age equa-tion of Kaufman et al (2008) does not support the indepen-

dent agendashdepth model in this area and is not applicable tosediments from the Yermak Plateau

Ages calculated with the globally calibrated age equationof Kaufman et al (2013) show the highest level of agreementwith sample ages when DL Asp values from N pachydermasamples are used Ages obtained using the globally calibratedage equation and DL Asp values in C neoteretis samplesconsistently appear younger (on average by 32 ) than theindependent ages as expected given the lower rate of AARin this species

The results highlight the need for further studies to testand explain the origin of the higher racemization rates in



foraminifera reported in previous studies from central Arc-tic sediments

Data availability The results of amino acid analysis of all sub-samples are available in the digital Supplement and archived at theWorld Data Service for Paleoclimatology (httpswwwncdcnoaagovpaleo-searchstudy27812 last access 15 November 2019West et al 2019)



Appendix A

Prior to alignment the kARMk time series and the δ18OLRstack were rescaled between minus1 and 1 to improve compara-bility The alignment function is then inferred at any pointon the sediment core depth scale by a linear interpolationbetween three proposed depthndashage models a starting nodean ending node and a perturbed node at a random locationbetween the starting and ending nodes The nodes strictlyfollow depthndashage paths that do not violate the principle ofsuperposition in order to ensure a monotonic relationshipbetween the depth of the sediment core and the age of theδ18OLR stack The algorithm determines the set of nodes thatgive a good alignment between the kARMk time series andthe δ18OLR stack as defined by a small residual standarddeviation and a high coefficient of correlation As an addi-tional condition we require that the gradient of the depthndashage alignment function should be reasonably smooth henceassuming that at our coring sites sedimentation rates have notundergone excessively rapid shifts or large fluctuations

The requirement of a good match between the kARMkrecords and the δ18OLR stack and a smooth alignment func-tion is here devised in a Bayesian formulation sensitive tospecification of a conjugate prior distribution and a conjugatelikelihood distribution of the alignment function The priorsspecify (i) the age distribution of the first and last kARMkvalues on the δ18OLR age scale (here set as non-informativeuniform ranging over 10 kyr) and (ii) the degree of autocor-relation of the sedimentation rates (here set as truncated uni-form between minus50 and +50 ) The latter dictates howmuch the accumulation rate of a particular depth depends onthe depth above it which in turn controls the gradient of thedepthndashage alignment function The likelihood distribution ofthe alignment function weighs the competing needs of smallstandard deviations of the data residuals and a high correla-tion between the kARMk and the δ18OLR values Finally aconditional posterior distribution of alignment functions pro-portional to the product of prior and likelihood is inferredCalculation of the posterior proceeds by sampling an initialvalue for each unknown parameter from the associated priordistributions using a reversible jump Markov chain MonteCarlo (MCMC) sampling (Vihola 2012) In the alignmentproblem posed here the automated MCMC method consistsof the following steps Starting from an initial age for thefirst and last kARMk values on the δ18OLR timescale thatdefines an initial match between the kARMk and δ18OLRtime series and by setting an initial value for the rate of sed-imentation it continues by

1 proposing a new ldquocandidaterdquo alignment function byadding a new depthndashage node at a random location be-tween the starting and ending nodes along the depthscale of the sediment core

2 accepting or rejecting the candidate alignment func-tion according to its posterior probability using the

MetropolisndashHastings algorithm (Metropolis et al 1953Hastings 1970) whereby the posterior probability ishigher for alignment functions that yield a closer match(ie a smaller residual standard deviation and a high co-efficient of correlation)

3 Repeat from step 1 for 106 iterations



Supplement The supplement related to this article is availableonline at httpsdoiorg105194gchron-1-53-2019-supplement

Author contributions GW analysed the data and prepared thepaper with contributions from all authors DSK contributed inter-pretation of AAR results FM provided age model tuning MF JMand JW contributed material chronological and oceanographic in-terpretation and MO coordinated the study and contributed palaeo-ceanographic interpretation

Competing interests The authors declare that they have no con-flict of interest

Acknowledgements We thank the Captain and the crew of RVPolarstern and the participants of the TRANSSIZ expedition for fa-cilitating data collection during PS92 Katherine Whitacre providedlaboratory support at the Amino Acid Geochronology LaboratoryNorthern Arizona University

Financial support This research has been supported by theSwedish Research Council (grant DNR-2016-05092) the NationalScience Foundation (grant NSF-1855381) and the Alfred-Wegener-Institut (grant AWI_PS92_00)

Review statement This paper was edited by Kirsty Penkman andreviewed by three anonymous referees

References

Alexanderson H Backman J Cronin T M Funder S Ingolfs-son O Jakobsson M Landvik J Y Loumlwemark L MangerudJ Maumlrz C and Moumlller P An Arctic perspective on dating Mid-Late Pleistocene environmental history Quaternary Sci Rev 929ndash31 httpsdoiorg101016jquascirev201309023 2014

Backman J Jakobsson M Loslashvlie R Polyak L andFebo L A Is the central Arctic Ocean a sedimentstarved basin Quaternary Sci Rev 23 1435ndash1454httpsdoiorg101016jquascirev200312005 2004

Backman J Jakobsson M Frank M Sangiorgi F BrinkhuisH Stickley C OrsquoRegan M Loslashvlie R Paumllike H Spof-forth D and Gattacecca J Age model and core-seismic in-tegration for the Cenozoic Arctic Coring Expedition sedimentsfrom the Lomonosov Ridge Paleoceanography 23 PA1S03httpshttpsdoiorg1010292007PA001476 2008

Bergsten H Recent benthic foraminifera of a transect fromthe North Pole to the Yermak Plateau eastern central ArcticOcean Mar Geol 119 251ndash267 httpsdoiorg1010160025-3227(94)90184-8 1994

Carstens J and Wefer G Recent distribution of planktonicforaminifera in the Nansen Basin Arctic Ocean Deep SeaResPart A 39 S507ndashS524 httpsdoiorg101016s0198-0149(06)80018-x 1992

Carstens J Hebbeln D and Wefer G Distribution of plankticforaminifera at the ice margin in the Arctic (Fram Strait) MarMicropaleontol 29 257ndash269 httpsdoiorg101016s0377-8398(96)00014-x 1997

Clark D Magnetic Reversals and Sedimenta-tion Rates in the Arctic Ocean GSA Bull81 3129ndash3134 httpsdoiorg1011300016-7606(1970)81[3129MRASRI]20CO2 1970

Goodfriend G A Brigham-Grette J and Miller G H Enhancedage resolution of the marine Quaternary record in the Arctic us-ing aspartic acid racemization dating of bivalve shells Quater-nary Res 45 176ndash187 httpsdoiorg101006qres199600181996

Greco M Jonkers L Kretschmer K Bijma J and Kucera MDepth habitat of the planktonic foraminifera Neogloboquadrinapachyderma in the northern high latitudes explained by sea-iceand chlorophyll concentrations Biogeosciences 16 3425ndash3437httpsdoiorg105194bg-16-3425-2019 2019

Haake F W and Pflaumann U W E Late Pleistoceneforaminiferal stratigraphy on the Voslashring Plateau Norwe-gian Sea Boreas 18 343ndash356 httpsdoiorg101111j1502-38851989tb00410x 1989

Hastings W K Monte Carlo sampling methods using Markovchains and their applications Biometrika 57 97ndash109httpsdoiorg101093biomet57197 1970

Hearty P J OrsquoLeary M J Kaufman D S Page M C andBright J Amino acid geochronology of individual foraminifer(Pulleniatina obliquiloculata) tests north Queensland marginAustralia a new approach to correlating and dating Quaternarytropical marine sediment cores Paleoceanography 19 PA4022httpsdoiorg1010292004PA001059 2004

Jakobsson M Loslashvlie R Arnold E M Backman J PolyakL Knutsen J O and Musatov E Pleistocene stratigraphyand paleoenvironmental variation from Lomonosov Ridge sed-iments central Arctic Ocean Global Planet Change 31 1ndash22httpsdoiorg101016s0921-8181(01)00110-2 2001

Jakobsson M Mayer L Coakley B Dowdeswell J A ForbesS Fridman B Hodnesdal H Noormets R Pedersen RRebesco M and Schenke H W The international bathymet-ric chart of the Arctic Ocean (IBCAO) version 30 GeophysRes Lett 39 L12609 httpsdoiorg1010292012GL0522192012

Jones E P Circulation in the arctic ocean Polar Res 20 139ndash146httpsdoiorg101111j1751-83692001tb00049x 2001

Kaufman D S Temperature sensitivity of asparticand glutamic acid racemization in the foraminiferaPulleniatina Quaternary Geochronol 1 188ndash207httpsdoiorg101016jquageo200606008 2006

Kaufman D S and Manley W F A new procedure for de-termining DL amino acid ratios in fossils using reverse phaseliquid chromatography Quaternary Sci Rev 17 987ndash1000httpsdoiorg101016s0277-3791(97)00086-3 1998

Kaufman D S Polyak L Adler R Channell J E andXuan C Dating late Quaternary planktonic foraminiferNeogloboquadrina pachyderma from the Arctic Ocean us-ing amino acid racemization Paleoceanography 23 PA3224httpshttpsdoiorg1010292008PA001618 2008

Kaufman D S Cooper K Behl R Billups K BrightJ Gardner K Hearty P Jakobsson M Mendes



I OrsquoLeary M and Polyak L Amino acid racem-ization in mono-specific foraminifera from Quaternarydeep-sea sediments Quaternary Geochronol 16 50ndash61httpsdoiorg101016jquageo201207006 2013

King K and Neville C Isoleucine epimerization for dat-ing marine sediments importance of analyzing monospe-cific foraminiferal samples Science 195 1333ndash1335httpsdoiorg101126science19542841333 1977

Kosnik M A and Kaufman D S Identifying outliers and assess-ing the accuracy of amino acid racemization measurements forgeochronology II Data screening Quaternary Geochronol 3328ndash341 httpsdoiorg101016jquageo200804001 2008

Kremer A Stein R Fahl K Ji Z Yang Z Wiers SMatthiessen J Forwick M Loumlwemark L OrsquoRegan M andChen J Changes in sea ice cover and ice sheet extent atthe Yermak Plateau during the last 160 ka ndash Reconstructionsfrom biomarker records Quaternary Sci Rev 182 93ndash108httpsdoiorg101016jquascirev201712016 2018

Lisiecki L E and Raymo M E A Pliocene-Pleistocene stack of57 globally distributed benthic δ18O records Paleoceanography20 PA1003 httpsdoiorg1010292004PA001071 2005

Lougheed B C Metcalfe B Ninnemann U S and WackerL Moving beyond the age-depth model paradigm in deep-sea palaeoclimate archives dual radiocarbon and stable iso-tope analysis on single foraminifera Clim Past 14 515ndash526httpsdoiorg105194cp-14-515-2018 2018

Macko S A and Aksu A E Amino acid epimeriza-tion in planktonic foraminifera suggests slow sedimentationrates for Alpha Ridge Arctic Ocean Nature 322 730ndash732httpsdoiorg101038322730a0 1986

Malinverno A Data report Monte Carlo correlation of sedimentrecords from core and downhole log measurements at SitesU1337 and U1338 (IODP Expedition 321) edited by Paumllike HLyle M Nishi H Raffi I Gamage K Klaus A and the Ex-pedition 320321 Scientists Proc IODP 320321 Tokyo (Inte-grated Ocean Drilling Program Management International Inc)httpsdoiorg102204iodpproc3203212072013 2013

Martinson D G Pisias N G Hays J D Imbrie J MooreT C and Shackleton N J Age dating and the orbital the-ory of the ice ages Development of a high-resolution 0 to300000-year chronostratigraphy 1 Quaternary Res 27 1ndash29httpsdoiorg1010160033-5894(87)90046-9 1987

Metropolis N Rosenbluth A W Rosenbluth M N TellerA H and Teller E Equation of state calculations byfast computing machines J Chem Phys 21 1087ndash1092httpsdoiorg10106311699114 1953

Muschitiello F Deglacial impact of the Scandinavian Ice Sheet onthe North Atlantic climate system (Doctoral dissertation Depart-ment of Geological Sciences) 2016

Muschitiello F Pausata F S Watson J E SmittenbergR H Salih A A Brooks S J Whitehouse N JKarlatou-Charalampopoulou A and Wohlfarth B Fennoscan-dian freshwater control on Greenland hydroclimate shifts atthe onset of the Younger Dryas Nat Commun 6 8939httpsdoiorg101038ncomms9939 2015

Muschitiello F DrsquoAndrea W J Schmittner A Heaton T JBalascio N L DeRoberts N Caffee M W Woodruff T EWelten K C Skinner L C and Simon M H Deep-watercirculation changes lead North Atlantic climate during deglacia-

tion Nat Commun 10 1272 httpsdoiorg101038s41467-019-09237-3 2019

Okay N and Crane K Thermal rejuvenation of theYermak Plateau Mar Geophys Res 15 243ndash263httpsdoiorg101007bf01982384 1993

OrsquoRegan M Moran K Backman J Jakobsson M SangiorgiF Brinkhuis H Pockalny R Skelton A Stickley C KoccedilN and Brumsack H J Mid-Cenozoic tectonic and paleoenvi-ronmental setting of the central Arctic Ocean Paleoceanography23 PA1S20 httpsdoiorg1010292007PA001559 2008

Pados T and Spielhagen R F Species distribution and depth habi-tat of recent planktic foraminifera in Fram Strait Arctic OceanPolar Res 33 22483 httpsdoiorg103402polarv33224832014

Polyak L and Solheim A Late-and postglacial environ-ments in the northern Barents Sea west of Franz JosefLand Polar Res 13 197ndash207 httpsdoiorg101111j1751-83691994tb00449x 1994

Sejrup H P Miller G H Brigham-Grette J Loslashvlie R andHopkins D Amino acid epimerization implies rapid sedi-mentation rates in Arctic Ocean cores Nature 310 772ndash775httpsdoiorg101038310772a0 1984

Sejrup H P and Haugen J E Amino acid diagenesis in the ma-rine bivalve Arctica islandica Linneacute from northwest Europeansites only time and temperature J Quaternary Sci 9 301ndash309httpsdoiorg101002jqs3390090402 1994

Shephard G E Wiers S Bazhenova E Peacuterez L F Mejiacutea LM Johansson C Jakobsson M and OrsquoRegan M A NorthPole thermal anomaly Evidence from new and existing heat flowmeasurements from the central Arctic Ocean J Geodyn 118166ndash181 httpsdoiorg101016jjog201801017 2018

Slubowska M A Koccedil N Rasmussen T L and Klitgaard-Kristensen D Changes in the flow of Atlantic water into theArctic Ocean since the last deglaciation evidence from the north-ern Svalbard continental margin 80 N Paleoceanography 20PA4014 httpsdoiorg1010292005PA001141 2005

Stein R The great challenges in Arctic Ocean paleoceanog-raphy IOP Conf Ser Earth Environ Sci 14 012001httpsdoiorg1010881755-1315141012001 2011

Vihola M Robust adaptive Metropolis algorithm with co-erced acceptance rate Stat Comput 22 997ndash1008httpsdoiorg101007s11222-011-9269-5 2012

Wehmiller J F Interlaboratory comparison of amino acid enan-tiomeric ratios in fossil Pleistocene mollusks Quaternary Res22 109ndash120 httpsdoiorg1010160033-5894(84)90010-31984

West G Kaufman D S Muschitiello F Forwick MMatthiessen J Wollenburg J and OrsquoRegan M YermakPlateau Arctic Ocean 227 ka Foraminifer Amino Acid Racem-ization Data NOAA available at httpswwwncdcnoaagovpaleo-searchstudy27812 last access 15 November 2019

Wiers S Snowball I OrsquoRegan M and Almqvist BLate Pleistocene Chronology of Sediments from the Yer-mak Plateau and Uncertainty in Dating Based on Geomag-netic Excursions Geochem Geophys Geosys 20 3289ndash3310httpsdoiorg1010292018gc007920 2019

Wohlfarth B Luoto T P Muschitiello F Vaumlliranta M BjoumlrckS Davies S M Kylander M Ljung K Reimer P Jand Smittenberg R H Climate and environment in south-



west Sweden 155ndash113 cal ka BP Boreas 47 687ndash710httpsdoiorg101111bor12310 2018

Wollenburg J E and Mackensen A Living benthic foraminifersfrom the central Arctic Ocean faunal composition stand-ing stock and diversity Mar Micropaleontol 34 153ndash185httpsdoiorg101016s0377-8398(98)00007-3 1998

Wollenburg J E Kuhnt W and Mackensen A Changes in Arc-tic Ocean paleoproductivity and hydrography during the last 145kyr the benthic foraminiferal record Paleoceanogr Paleocl 1665ndash77 2001

Wollenburg J E Knies J and Mackensen A High-resolutionpaleoproductivity fluctuations during the past 24 kyr as indicatedby benthic foraminifera in the marginal Arctic Ocean Palaeo-geogr Palaeocl 204 209ndash238 httpsdoiorg101016s0031-0182(03)00726-0 2004

Xuan C Channell J E Polyak L and Darby D APaleomagnetism of Quaternary sediments from LomonosovRidge and Yermak Plateau implications for age modelsin the Arctic Ocean Quaternary Sci Rev 32 48ndash63httpsdoiorg101016jquascirev201111015 2012


Abstract

Introduction

Materials and methods

Sample material and study area

Sample ages and quantification of age model uncertainties

Sample preparation and analytical procedure

Results

Discussion

Relation between DL values and depth and inter-specific differences

Relation between DL values and sample age

Palaeotemperature and other possible effects

Comparing AAR age models

Conclusions

Data availability

Appendix A

Supplement

Author contributions

Competing interests

Acknowledgements

Financial support

Review statement

References


sample size and low temporal resolution ndash of these early stud-ies the application of AAR chronology for Arctic Ocean sed-iments decreased in the following years However major de-velopments in AAR dating such as the increased use of themore rapidly racemizing aspartic acid over isoleucine (egGoodfriend et al 1996) and the application of the reversephase liquid chromatography to determine amino acid DLvalues in small samples (Kaufman and Manley 1998) cou-pled with improvements in Arctic Ocean sediment chronos-tratigraphies (eg Jakobsson et al 2001 Backman et al2008 OrsquoRegan et al 2008) have reignited the interest inAAR studies during the past decade

Kaufman et al (2008) related the extent of AAR in thepolar foraminifera species Neogloboquadrina pachydermacollected from various parts of the Arctic Ocean (MendeleevRidge Lomonosov Ridge Northwind Ridge) to sedimentages by developing a calibrated age equation for the past150 kyr This was followed by the study of Kaufman etal (2013) who reported the results of AAR analysis of mul-tiple foraminifera species from sediment cores taken in theArctic Atlantic and Pacific oceans and defined a generalrate of racemization at deep-sea sites Both models relied oncalibration of the rate of racemization of aspartic (Asp) andglutamic (Glu) acids by independent dating methods (such asradiocarbon dating and correlations with the orbitally tunedstacked oxygen isotope record Martinson et al 1987) yetfor the past 150 kyr these two age equations produced sig-nificantly different ages for a given DL value The globalcompilation of AAR showed that the rate of racemization inthe Arctic Ocean was higher than expected in samples of Npachyderma older than 35 ka and this could not be explainedby differences in the rate of AAR in N pachyderma rela-tive to other species It remains to be seen whether the ob-served rates of racemization are anomalously high comparedto other ocean basins or alternatively the age constraints forthe studied Arctic Ocean cores are inaccurate Sedimentarysequences with robust age control and satisfactory preserva-tion of carbonate microfossils from the Arctic are required tofurther investigate this question

Here we present the results of amino acid racemiza-tion analysis of planktic (N pachyderma) and benthic (Cneoteretis) foraminifera from samples of three well-datedsediment cores from the Yermak Plateau The rates of racem-ization in the two species are compared and the Arctic-specific (Kaufman et al 2008) and global (Kaufman et al2013) calibrated age equations are used to estimate the agesof the samples The calculated ages are compared with thosederived from age models constructed by combining 14C dat-ing oxygen isotope stratigraphy and the correlation of envi-ronmental magnetic properties to the global benthic oxygenisotope record

Figure 1 Locations of the studied cores and water depth at coresites Basemap IBCAO Jakobsson et al (2012)

2 Materials and methods

21 Sample material and study area

The analysed foraminifera samples were taken from threesediment cores (PS920039-2 PS920045-2 and PS9254-1)recovered from the Yermak Plateau during the TRANSSIZ(TRansitions in the Arctic Seasonal Sea Ice Zone) expedi-tion (19 May to 28 June 2015) on the research icebreaker RVPolarstern (PS92) (Table 1 and Fig 1)

The Yermak Plateau located north of the Svalbardarchipelago lies at the gateway to the central Arctic Oceanin an area characterized by the interaction between Atlanticand Arctic waters The cold surface layer is dominated by thepolar foraminifer N pachyderma which has highest abun-dances in the top 200 m (Carstens and Wefer 1992 Carstenset al 1997 Pados and Spielhagen 2014 Greco et al 2019)Cassidulina neoteretis is one of the dominant benthic speciesin sediments from the Yermak Plateau (Bergsten 1994 Wol-lenburg and Mackensen 1998 Wollenburg et al 2004)and its abundance is often associated with the presence ofAtlantic water (Polyak and Solheim 1994 Slubowska etal 2005)

The cores were recovered from water depths between 915and 1464 m ndash in the Upper Polar Deep Water ndash far be-low the core of the warmer Atlantic layer which typicallysits between depths of 200 and 500 m (Jones 2001) Po-tential temperature and salinity profiles from nearby CTD(conductivityndashtemperaturendashdepth) stations illustrate the re-





















3 Results





























4 Discussion

















































5 Conclusions












Appendix A















References


























































Abstract

Introduction





Results

Discussion





Conclusions

Data availability

Appendix A

Supplement


Competing interests

Acknowledgements

Financial support

Review statement

References




















3 Results





























4 Discussion

















































5 Conclusions












Appendix A















References


























































Abstract

Introduction





Results

Discussion





Conclusions

Data availability

Appendix A

Supplement


Competing interests

Acknowledgements

Financial support

Review statement

References



























4 Discussion

















































5 Conclusions












Appendix A















References


























































Abstract

Introduction





Results

Discussion





Conclusions

Data availability

Appendix A

Supplement


Competing interests

Acknowledgements

Financial support

Review statement

References












































5 Conclusions












Appendix A















References


























































Abstract

Introduction





Results

Discussion





Conclusions

Data availability

Appendix A

Supplement


Competing interests

Acknowledgements

Financial support

Review statement

References






Appendix A















References


























































Abstract

Introduction





Results

Discussion





Conclusions

Data availability

Appendix A

Supplement


Competing interests

Acknowledgements

Financial support

Review statement

References








References


























































Abstract

Introduction





Results

Discussion





Conclusions

Data availability

Appendix A

Supplement


Competing interests

Acknowledgements

Financial support

Review statement

References






































Abstract

Introduction





Results

Discussion





Conclusions

Data availability

Appendix A

Supplement


Competing interests

Acknowledgements

Financial support

Review statement

References

Amino acid racemization in Quaternary foraminifera from the … · 2020-02-28 · 2008; O’Regan...

Documents

Transcript of Amino acid racemization in Quaternary foraminifera from the … · 2020-02-28 · 2008; O’Regan...